Glioblastoma multiforme (GBM) is an aggressive and incurable cancer. The dire prognosis for this most prevalent form of malignant brain tumor is unchanged over 30 years. The Cancer Genome Atlas (TCGA) project has genomically surveyed several hundred human GBM samples and identified a small spectrum of genes that are frequently somatically mutated and include the tumor suppressors NF1, p53 and Pten. We have generated and studied mouse models that spontaneously form classic GBM with 100% incidence. The models are based on somatic mutation of NF1, p53, and Pten. The full penetrance of these tumors has allowed us to probe early pre-tumorigenic events and to identify the adult neural stem cell (NSC) niche as the prevalent source of these tumors. In the previous funding period, we exploited these models to demonstrate a stem cell origin of these tumors and exclude differentiated brain cells as a source. We probed transcriptional profiles to identify glioma-specific upregulation of HoxA genes and examined their function. We developed tools to directly test and demonstrate in vivo the existence of a hierarchical population of cancer stem cells (CSCs) by direct lineage tracing in spontaneous gliomas without resorting to ex vivo transplantation and/or culturing. We unveiled a novel GBM that arises from a different cell of origin. Finally, we developed a novel model of metastatic medulloblastoma that depends on Dicer mutation and is being developed as a separate research initiative. The present proposal embodies two Specific Aims. Aim 1, entitled: Cancer stem cells in glioma: identification, isolation and characterization. pursues our recent success in transgenically labeling an endogenous glioma cell population that is responsible for tumor recurrence after chemotherapy. We have devised an advanced NSC- and CSC-specific transgene that will co-express Cre recombinase, eGFP, and human diphtheria toxin receptor. This transgene, when bred into the Nf1;p53;Pten floxed background, will permit more precise identification and study of the relatively quiescent CSC population. It will allow purification for quantitative RNA seq analysis i comparison to the non-CSC tumor population, and to the progenitor NSC population. These studies will provide novel information and insight into these newly discovered CSCs. Aim 2 will examine whether neural progenitor cells (NPCs) and oligodendrocyte progenitor cells (OPCs) give rise to distinct forms of glioma and GBM. We provide evidence that NPCs and OPCs can, by mutations in NF1, p53 and Pten, give rise to gliomas. Mutant NPCs give rise to tumors very similar to the tumors arising in the stem cell compartments whereas OPCs generate an independent form of GBM with distinct molecular, and developmental characteristics. We propose to study these novel tumors in depth since, to the pathologist, they belong to a single GBM pool. Our detailed studies of these tumors may inform human tumor analysis and serve to separate these OPC GBMs that may have different properties, including response to therapies. Genomic, functional, and CSC studies will be pursued for these novel brain tumors.